Update of the mutation spectrum and clinical correlations of over 360 mutations in eight genes that underlie the neuronal ceroid lipofuscinoses

Authors

  • Maria Kousi,

    1. Folkhälsan Institute of Genetics, Department of Medical Genetics, Haartman Institute, and Neuroscience Center, University of Helsinki, Finland
    Search for more papers by this author
  • Anna-Elina Lehesjoki,

    Corresponding author
    1. Folkhälsan Institute of Genetics, Department of Medical Genetics, Haartman Institute, and Neuroscience Center, University of Helsinki, Finland
    • Folkhälsan Institute of Genetics, Biomedicum Helsinki, P.O. Box 63 (Haartmaninkatu 8), 00014 University of Helsinki, Finland
    Search for more papers by this author
  • Sara E. Mole

    1. MRC Laboratory for Molecular Cell Biology, Molecular Medicine Unit, UCL Institute of Child Health, and Department of Genetics, Environment and Evolution, University College London, United Kingdom
    Search for more papers by this author

  • Communicated by Christine Van Broeckhoven

Abstract

The neuronal ceroid lipofuscinoses (NCLs) are clinically and genetically heterogeneous neurodegenerative disorders. Most are autosomal recessively inherited. Clinical features include a variable age of onset, motor and mental decline, epilepsy, visual loss, and premature death. Mutations in eight genes (PPT1/CLN1, TPP1/CLN2, CLN3, CLN5, CLN6, MFSD8/CLN7, CLN8) have been identified and several more are predicted to exist, including two provisionally named CLN4 and CLN9. Despite excessive in vitro and in vivo studies, the precise functions of the NCL proteins and the disease mechanisms remain elusive. To date 365 NCL-causing mutations are known, with 91 novel disease-causing mutations reported. These are reviewed with an emphasis on their complex correlation to phenotypes. Different mutations within the NCL spectrum can cause variable disease severity. The NCLs exemplify both phenotypic convergence or mimicry and phenotypic divergence. For example, mutations in CLN5, CLN6, MFSD8, or CLN8 can underlie the clinically similar late infantile variant NCL disease. Phenotypic divergence is exemplified by different CLN8 mutations giving rise to two very different diseases, the mild CLN8 disease, EPMR (progressive epilepsy with mental retardation), and the more severe CLN8 disease, late infantile variant. The increase in the genetic understanding of the NCLs has led to improved diagnostic approaches, and the recent proposal of a new nomenclature. Hum Mutat 33:42–63, 2012. © 2011 Wiley Periodicals, Inc.

Introduction

The neuronal ceroid lipofuscinoses (NCLs) comprise a group of progressive neurodegenerative disorders mainly affecting children. Most types of NCL show autosomal recessive inheritance. The main clinical symptoms involve mental and motor deterioration, epilepsy, visual loss, ataxia, and a reduced life span [Haltia, 2003]. The age of onset can be variable. The unifying finding in NCL patients is the accumulation, in neurons and many other cell types, of autofluorescent storage material revealed upon electron microscopy (EM) examination of tissue material [Haltia, 2003]. Historically, classification has been based on age of onset together with clinical presentation, ultrastructural morphology of the storage material, and, more recently, genetic defect. Patients were grouped in one of the six basic NCL subtypes, consisting of congenital (CLN10), infantile (CLN1), late infantile (CLN2), variant late infantile (CLN5, CLN6, CLN7, and CLN8), juvenile (CLN3), and adult (CLN4 or Kufs disease) NCL [Mole et al., 2005]. Since there can be very wide variation in age of onset and disease progression, in many cases genetic testing and identification of the causative mutations are required in order to be able to unequivocally place a patient within the correct subtype. For this reason, a novel NCL nomenclature has been proposed that is genetically based but still takes into account key clinical and pathological features. In the novel nomenclature the gene mutated denotes the specific subtype, and is followed by the clinical presentation based on symptomatology, age of onset, and EM findings, for example, CLN3 disease, juvenile (Table 1). The data presented in this update will follow this new nomenclature.

Table 1. Summary of New and Former Nomenclature of the Neuronal Ceroid Lipofuscinoses
GeneMIM #EponymDiseaseFormer abbreviated name
  1. CNCL, congenital ceroid-lipofuscinosis; INCL, infantile ceroid-lipofuscinosis; LINCL, late infantile ceroid-lipofuscinosis; vLINCL, variant late infantile ceroid-lipofuscinosis; JNCL juvenile ceroid-lipofuscinosis; ANCL, adult onset neuronal ceroid-lipofuscinosis; EPMR, progressive epilepsy with mental retardation.

PPT1/CLN1256730Haltia-SantavuoriCLN1 disease, classic infantile,INCL
 CLN1 disease, late infantile, 
 CLN1 disease, juvenile,JNCL/GROD
 CLN1 disease, adult 
TPP1/CLN2204500Janský-BielschowskyCLN2 disease, classic late infantileLINCL
 CLN2 disease, juvenile 
 CLN2 disease, infantile 
CLN3204200Spielmeyer-SjögrenCLN3 disease, classic juvenileJNCL
 CLN3 disease, protracted 
 CLN3 disease, infantile 
CLN5256731Finnish variant late infantileCLN5 disease, late infantile variantvLINCL
 CLN5 disease, juvenile 
 CLN5 disease, adult 
 CLN5 disease, infantile 
CLN6601780Lake-Cavanagh early juvenile variant or Indian variant late infantileCLN6 disease, late infantile variantvLINCL
 CLN6 disease, adult or Kufs type AKufs type A
MFSD8/CLN7610951Turkish variant late infantileCLN7 disease, late infantile variantvLINCL
 CLN7 disease, juvenile 
CLN8600143variant late infantileCLN8 disease, late infantile variantvLINCL
 and Northern epilepsy/EPMRCLN8 disease, EPMREPMR
CTSD/CLN10610127CongenitalCLN10 disease, congenitalCNCL
 CLN10 disease, juvenile 
?609055Juvenile variantCLN9 disease, juvenile variantvJNCL
?204300Kufs (type A or type B)adult (autosomal recessive) NCL disease or KufsANCL
?162350Parryadult (autosomal dominant) NCL disease or ParryANCL

In addition to clinical heterogeneity, the NCLs are also characterized by genetic heterogeneity. The eight NCL-associated genes known to date have mainly been identified using molecular genetic methods [PPT1, Vesa et al., 1995; CLN3, The International Batten Disease Consortium, 1995; CLN5, Savukoski et al., 1998; CLN6, Gao et al., 2002; Wheeler et al., 2002; MFSD8, Siintola et al., 2007; CLN8, Ranta et al., 1999; CTSD; Siintola et al., 2006b; Steinfeld et al., 2006] or a biochemical approach [TPP1, Sleat et al., 1997]. Prior to this update, 274 mutations had been found to affect the eight NCL genes (described in the literature and listed in the web-based NCL Mutation Database www.ucl.ac.uk/ncl/mutation). Although NCL genetic defects in each of the known genes have been shown to have a worldwide distribution, some mutations seem to be more common than others in specific populations, most probably having been established by founder effects [reviewed by Mole et al., 2005]. Nevertheless, the genetic spectrum underlying this group of disorders remains incomplete.

Four of the NCL genes encode soluble lysosomal proteins (PPT1, TPP1, CTSD, CLN5) and four encode transmembrane proteins that reside in either lysosomes (CLN3, MFSD8) or the endoplasmic reticulum (ER) (CLN6, CLN8) [reviewed by Jalanko and Braulke, 2009]. Some NCL genes are highly conserved suggesting that they play a fundamental role in eukaryotic cells. Expression of NCL proteins is ubiquitous, although the neurons are the cell-type that is predominantly affected when the proteins are defective.

Naturally occurring NCL animal models that recapitulate key clinical features of the NCL phenotype exist in dog, sheep, cow, and mouse (http://www.ucl.ac.uk/ncl/animal.shtml). In addition to these, several experimental animal models have been and are being developed (http://www.ucl.ac.uk/ncl/animal.shtml). Among the vertebrate experimental animals, mouse models exist for all subtypes except CLN7 disease [reviewed by Cooper et al., 2006], and zebrafish systems are currently being developed (www.ucl.ac.uk/ncl/animals). The invertebrate model organisms include yeast (Saccharomyces cerevisiae and Schizosaccharomyces pombe), nematode worm (Caenorhabditis elegans), and fruitfly (Drosophila melanogaster) models [reviewed by Phillips et al., 2006]. To date, despite extensive study in both mammalian and model systems, the biological function of the individual NCL proteins and their associated disease mechanisms remains elusive.

The purpose of this mutation update is to summarize the genetic variation that occurs in the NCL-associated genes, with emphasis on genotype–phenotype correlations. A total of 91 novel mutations are reported, raising the total number of NCL causing changes to 365 (Table 2). In this update, each gene is considered in turn. Each NCL-causing mutation is presented both by gene and by patient/family in the Supporting Information tables (Supp. Tables S1–S16). This aims in facilitating the construction of a complete genetic profile for all NCL and other disease associated genes screened in each person diagnosed with NCL disease. The latter comprises an essential step toward identifying modifiers that underlie inter- and intrafamilial variation, and understanding the beneficial and harmful effects of future therapeutic attempts.

Table 2. Novel Mutations Identified in NCL-Associated Genes
GeneExon/IntronDNA variantPredicted/Demonstrated effect on the proteinCountry of origin
  1. Nucleotide numbering is based on GenBank reference sequences NM_000310.3 for PPT1, NM_000391.3 for TPP1, NM_000086.2 for CLN3, NM_006493.2 for CLN5, NM_017882.2 for CLN6, NM_152778.2 for MFSD8, and NM_018941.3 for CLN8.

PPT1Exon 1c.114G>Ap.Trp38XUK
PPT1Exon 1c.114delGp.Trp38CysfsX12UK
PPT1Intron 2c.235-3T>Csplice defectItaly
PPT1Intron 3c.363-4G>Asplice defectTurkey
PPT1Exon 4c.413C>Tp.Ser138LeuTurkey
PPT1Intron 5c.536+2T>Csplice defectUK
PPT1Exon 6c.538dupCp.Leu180ProfsX9Turkey
PPT1Exon 6c.558G>Ap.Trp186XUK
PPT1Exon 6c.560A>Gp.His187ArgUK
PPT1Exon 6c.566C>Gp.Pro189ArgTurkey
PPT1Exon 7c.683T>Gp.Val228GlyThe Netherlands
PPT1Exon 9c.886T>Cp.Trp296ArgUK
PPT1Exon 9c.914T>Cp.Leu305ProFinland
TPPIIntron 1c.18-3C>Gsplice defectUK
TPPIExon 2c.37dupCp.Leu13ProfsX32UK
TPPIExon 3c.184T>Ap.Ser62ThrTurkey
TPPIExon 4c.237C>Gp.Tyr79XCzech Republic
TPP1Intron 4c.381-1G>CSplice defect/frameshiftCanada
TPPIExon 5c.497dupAp.His166GlnfsX22Turkey
TPPIExon 6c.625T>Cp.Tyr209HisUK
TPPIExon 6c.640C>Tp.Gln214XItaly
TPPIExon 7c.790C>Tp.Gln264XItaly
TPPIExon 7c.797G>Ap.Arg266GlnSpain
TPPIExon 7c.822_837delp.Leu275XUK
TPPIExon 8c.1016G>Ap.Arg339GlnSpain
TPPIExon 8c.1062delGp.Leu355SerfsX72Italy
TPPIIntron 8c.1075+2T>Csplice defectThe Netherlands
TPP1Exon 10c.1146C>Gp.Ser382ArgSpain
TPPIExon 11c.1343C>Tp.Ala448ValTurkey
TPPIIntron 11c.1425+1G>Csplice defectBelgium
TPPIExon 12c.1497delTp.Gly501AlafsX18Turkey
TPPIExon 12c.1501G>Tp.Gly501CysTurkey
TPPIExon 12c.1510A>Tp.Asn504TyrUK
TPPIExon 12c.1547_1548delTTp.Phe516XTurkey
TPPIIntron 12c.1551+5_1551+6delGTinsTASplice defectPortugal
TPPIExon 13c.1642T>Cp.Trp548ArgSlovakia
CLN3Exon 1c.1A>Cp.Met1?The Netherlands
CLN3Exon 2c.105G>Ap.Trp35XGermany, USA
CLN3Intron 2c.125+5G>ASplice defectBelgium
CLN3Intron 2c.126-1G>ASplice defectTurkey
CLN3Intron 3c.222+5G>CSplice defectGermany
CLN3Exon 4c.233_234insGp.Thr80AsnfsX12Turkey
CLN3Exon 6c.400T>Cp.Cys134ArgItaly
CLN3Intron 6c.461-1G>CSplice defectCanada
CLN3Exon 8c.560G>Cp.Gly187AlaSpain, The Netherlands
CLN3Exon 8c.565G>Cp.Gly189ArgPortugal
CLN3Exon 12c.954_962+18del27p.Leu313_Trp321del / splice defectSweden, Norway, The Netherlands
CLN3Exon 12-Intron 12c.963-1G>TSplice defectUK
CLN3Exon 14c.1195G>Tp.Glu399XItaly
CLN5Exon 1c.61C>Tp.Pro21SerTurkey
CLN5Exon 1c.223T>Cp.Trp75ArgTurkey
CLN5Exon 2c.433C>Tp.Arg145XUK
CLN5Exon 3c.524T>Gp.Leu175XTurkey
CLN5Exon 3c.593T>Cp.Leu198ProTurkey
CLN5Exon 3c.613C>Tp.Pro205SerCanada, Qatar
CLN5Exon 3c.619T>Cp.Trp207ArgUK
CLN5Exon 4c.726C>Ap.Asn242LysUK
CLN5Exon 4c.955_970del16p.Gly319PhefsX12UK
CLN6Exon 1c.34G>Ap.Ala12ThrTurkey
CLN6Exon 1c.49G>Ap.Gly17SerTurkey
CLN6Exon 3c.251delAp.Tyr84SerfsX32China
CLN6Exon 3c.270C>Gp.Asn90LysIndia
CLN6Exon 4c.311C>Tp.Ser104PheCzech Republic
CLN6Exon 4c.445C>Tp.Arg149CysCanada
CLN6Intron 4c.486+8C>TSplice defectArgentina, UK, Canada
CLN6Exon 5c.506T>Cp.Leu169ProCzech Republic
CLN6Exon 5c.516T>Ap.Tyr172XTurkey
CLN6Exon 6c.557T>Cp.Phe186SerTurkey
CLN6Exon 7c.700T>Cp.Phe234LeuItaly
CLN6Exon 7c.727delGp.Ala243ProfsX26Italy/Lybia
CLN6Exon 7c.755G>Ap.Arg252HisArgentina, UK
CLN6Exon 7c.775G>Ap.Gly259SerIndia
CLN6Exon 7c.889C>Ap.Pro297ThrPakistan
MFSD8Exon 5c.259C>Tp.Gln87XCanada
MFSD8Exon 6c.479C>Ap.Thr160AsnTurkey
MFSD8Exon 6c.479C>Tp.Thr160IleCook Islands
MFSD8Intron 6c.554-1G>CSplice defectRomania
MFSD8Intron 8c.754+1G>ASplice defectTurkey
MFSD8Exon 13c.1373C>Ap.Thr458LysRomania
MFSD8Exon 13c.1394G>Ap.Arg465GlnTurkey
MFSD8Exon 13c.1408A>Gp.Met470ValTurkey
MFSD8Exon 13c.1420C>Tp.Gln474XTurkey
CLN8Exon 2c.209G>Ap.Arg70HisIndia, New Zealand
CLN8Exon 2c.227A>Gp.Gln76ArgTurkey
CLN8Exon 2c.320T>Gp.Ile107SerTurkey
CLN8Exon 2c.374A>Gp.Asn125SerTurkey
CLN8Exon 2c.415C>Tp.His139TyrNew Zealand
CLN8Exon 3c.637_639delTGGp.Trp213delTurkey
CLN8Exon 3c.661G>Ap.Gly221SerTurkey
CLN8Exon 3c.685C>Gp.Pro229AlaMexico, Argentina
CLN8Exon 3c.806A>Tp.Glu269ValTurkey

Mutation Nomenclature

The mutation nomenclature used in this update follows the guidelines indicated by Human Genome Variation Society (HGVS) [den Dunnen and Antonarakis, 2000]. Nucleotide numbering is based on GenBank reference sequences NM_000310.3 for PPT1, NM_000391.3 for TPP1, NM_000086.2 for CLN3, NM_006493.2 for CLN5, NM_017882.2 for CLN6, NM_152778.2 for MFSD8, NM_018941.3 for CLN8, and NM_001909.4 for CTSD. In all genes presented here, nucleotide numbering reflects cDNA numbering with position +1 corresponding to the A of the ATG translation initiation codon (in PPT1 NM_000310.3 nucleotide 233, in TPP1 NM_000391.3 nucleotide 62, in CLN3 NM_000086.2 nucleotide 324, in CLN5 NM_006493.2 nucleotide 29, in CLN6 NM_017882.2 nucleotide 159, in MFSD8 NM_152778.2 nucleotide 164, in CLN8 NM_018941.3 nucleotide 306, and in CTSD NM_001909.4 nucleotide 134, respectively correspond to +1). Mutation descriptions on the protein level consider the initiator methionine as codon 1 and have been checked using the Mutalyzer program (http://www.LOVD.nl/mutalyzer/).

CLN1

CLN1 Disease, Clinical Spectrum

Children affected with CLN1 disease, classic infantile (MIM# 256730), usually develop normally until 10–18 months of age [Santavuori et al., 2000]. In some cases, microcephaly is observed as early as the age of 5 months [Santavuori et al., 2000]. The majority of children show a decline in normal development during the second year of life. In CLN1 disease, classic infantile, symptoms include deterioration of motor skills, developmental delay, seizures, speech, and visual failure. Death usually occurs between 6 and 15 years of age [Santavuori, 1988].

In addition to the classic clinical phenotype, CLN1 disease has also been associated with phenotypes of later onset. These include CLN1 disease, late infantile, juvenile, and adult (Table 1). Association of PPT1 mutations with the CLN1 disease, late infantile phenotype, where disease onset is between 1.5 and 3 years has been reported in 13 cases [Bonsignore et al., 2006; Das et al., 1998; Simonati et al., 2009; Waliany et al., 2000; Wisniewski et al., 1998a]. CLN1 disease, juvenile, is very similar clinically to that of CLN3 disease, classic juvenile, differing by presenting with learning rather than visual difficulties, showing regression of acquired skills at earlier age, and having no intracellular accumulation of subunit c of mitochondrial ATP-synthase and no vacuolated lymphocytes in the patients' peripheral blood [Mitchison et al., 1998; Mueller and Coovadia, 2010; Pérez-Poyato et al., 2011]. Finally, three CLN1 disease, adult cases have been described, including one with onset before 20 years [Ramandan et al., 2007] and two sisters from France that presented at ages 31 and 38 years [van Diggelen et al., 2001].

CLN1 disease always correlates with the intracellular accumulation of granular osmiophilic deposits (GROD) consisting mainly of sphingolipid-activator proteins A and D [Carpenter et al., 1973; Das et al., 1998; Ramandan et al., 2007; Santavuori et al., 2000; va Diggelen et al., 2001; Wisniewski et al., 1998a].

PPT1 Gene and Protein

PPT1 (NM_000310.3) maps to chromosome 1p32 and encodes a soluble long fatty acid hydrolase, palmitoyl-protein thioesterase 1 (PPT1) [Järvelä et al., 1991; Vesa et al., 1995]. PPT1 is predicted to be composed of 306 amino acids, containing a 25 amino acid signal peptide and three N-linked glycosylation sites [Schriner and Hofmann, 1996]. Topologically, PPT1 is a globular enzyme that tends to oligomerize and contains an α/β-hydrolase fold with a catalytic triad of Ser115-Asp233-His289 [Bellizzi et al., 2000; Hofmann et al., 2002; Lyly et al., 2007].

The biochemical function of PPT1 involves removal of fatty acyl groups from modified cysteines in proteins [Camp and Hofmann, 1993]. Although the specific substrate(s) of PPT1 remain elusive, acyl-CoAs and cysteine residues of lipid-modified proteins can be substrates in vitro, and fatty acylated proteins are substrates in vivo [Lu et al., 1996]. PPT1 functions in the lysosomes, although in neuronal cells a colocalization with synaptic vesicles and synaptosomes rather than with lysosomes is proposed [Hellsten et al., 1996; Lyly et al., 2007]. Trafficking of PPT1 is mediated either by the mannose-6-phosphate (Man6P) receptor-mediated pathway or by an alternative yet undefined pathway [Hellsten et al., 1996; Lyly et al., 2007; Verkruyse and Hofmann, 1996].

PPT1 Mutation Spectrum

So far 61 PPT1 mutations (13 novel reported here and 48 previously reported) have been identified (Table 2, Fig. 1, and Supp. Tables S1 and S2). These consist of 27 missense, 11 nonsense, 10 splice-site affecting, four insertions, six deletions, one insertion-deletion, one that could be either missense or splice-site affecting, and one mutation affecting the initiator methionine. The residues Trp38, Cys152, Val181, and Trp296 are affected by more than one mutations (Fig. 1). The pathogenic role of the PPT1 mutations is supported by the resultant reduced enzyme activity [Das et al., 2001; Mitchison et al., 1998; Simonati et al., 2009; Sleat et al., 2001].

Figure 1.

Schematic representation of the PPT1 gene and the PPT1 protein showing the relative position of the mutations. The coding exons of PPT1 and the PPT1 protein are shown in scale as purple boxes. The untranslated exonic fragments and the introns are shown in gray boxes and lines, respectively, and are not in scale. The residues that participate in the formation of the catalytic triad of PPT1 are depicted as green circles. The residues at which N-glycosylation occurs are shown as red circles. The signal peptide is represented with a yellow square. The missense mutations are shown in the upper and other mutations in the lower part of the gene and the protein. The mutations previously reported in the literature are given in black color. The novel mutations reported here are shown in red. Nucleotide numbering is based on GenBank reference sequence NM_000310.3.

Among the 61 recognized mutations affecting PPT1, three are particularly common. The missense mutation p.Arg122Trp, affecting a conserved amino acid, occurs in all CLN1 disease, infantile, patients originating from Finland, and exemplifies a founder effect [Vesa et al., 1995]. A second founder effect is likely in the Scottish population where the p.Thr75Pro mutation, associated with CLN1 disease juvenile, is particularly common [Munroe et al., 1998]. Finally, the most commonly occurring mutation worldwide is the nonsense mutation p.Arg151X, which is encountered in approximately 40% of non-Finnish PPT1-affected alleles [Das et al., 1998; Mitchison et al., 1998; Mole et al., 2001; Munroe et al., 1998].

Genotype–Phenotype Correlations

PPT1 mutations can be associated with disease onset ranging from infantile to adult (Table 3). Inheritance of any combination of nonsense or frameshift alleles results in an infantile onset phenotype with disease onset at less than 2 years and rapid progression [Das et al., 1998]. This is most probably due to the loss of the p.His289 residue of the enzyme's catalytic triad [Das et al., 2001], causing very low levels of detectable enzyme activity. The identified missense mutations produce a spectrum of phenotypes from infantile to adult onset, or a more protracted clinical course. PPT1 missense mutations are believed to exert their deleterious effects by causing trapping of the mutant proteins in the ER where they are targeted for ER-mediated degradation (due to defective recognition of the polypeptides by the Man6P receptor) [Das et al., 2001]. The milder effect of some missense mutations could be explained in the light of new findings that suggest the use of alternative sorting pathways for PPT1 in addition to the Man6P receptor-mediated pathway, or alternatively by the observation that mutant PPT1 molecules show a higher degree of oligomerization, possibly as an attempt to regulate activation and/or transport of the enzyme [Lyly et al., 2007].

Table 3. Mutations Causing Atypical Phenotypes
GeneExon/IntronDNA variantPredicted/Demonstrated effect on the proteinPhenotype
  1. Nucleotide numbering is based on GenBank reference sequences NM_000310.3 for PPT1, NM_000391.3 for TPP1, NM_000086.2 for CLN3, NM_006493.2 for CLN5, NM_017882.2 for CLN6, NM_152778.2 for MFSD8, NM_018941.3 for CLN8, and NM_001909.3 for CTSD.

PPT1Exon 1c.3G>AInefficient initiation (p.Met1?)CLN1 disease, late infantile
 CLN1 disease, juvenile
PPT1Intron 1c.125-15T>GSplice defect/FrameshictCLN1 disease, late infantile
PPT1Exon 2c.134G>Ap.Cys45TyrCLN1 disease, adult
PPT1Exon 2c.223A>Cp.Thr75ProCLN1 disease, juvenile
PPT1Intron 2c.235-3T>CSplice defectCLN1 disease, juvenile
PPT1Exon 3c.236A>Gp.Asp79GlyCLN1 disease, juvenile
PPT1Exon 3c.322G>Cp.Gly108ArgCLN1 disease, adult
PPT1Exon 3c.325T>Gp.Tyr109AspCLN1 disease, late infantile
PPT1Intron 3c.363-3T>GSplice defect/FrameshiftCLN1 disease, juvenile
PPT1Exon 5c.451C>Tp.Arg151XCLN1 disease, late infantile
PPT1Exon 5c.455G>Ap.Cys152TyrCLN1 disease, juvenile
PPT1Exon 5c.490C>Tp.Arg164XCLN1 disease, late infantile
PPT1Exon 5c.529C>Gp.Gln177GluCLN1 disease, late infantile
PPT1Intron 5c.536+1G>ASplice defect/FrameshiftCLN1 disease, juvenile
PPT1Exon 6c.541G>Ap.Val181MetCLN1 disease, late infantile
 CLN1 disease, juvenile
PPT1Exon 7c.656T>Ap.Leu219GlnCLN1 disease, juvenile
PPT1Exon 7c.665T>Cp.Leu222ProCLN1 disease, late infantile
 CLN1 disease, juvenile
PPT1Exon 8c.739T>Cp.Tyr247HisCLN1 disease, juvenile
PPT1Exon 8c.749G>Tp.Gly250ValCLN1 disease, juvenile
TPPIExon 4c.380G>Ap.Arg127GlnCLN2 disease, late infantile, protracted course
TPP1Exon 6c.622C>Tp.Arg208XCLN2 disease, juvenile
 CLN2 disease, infantile
TPP1Intron 5c.509-1G>CSplice defect/FrameshiftCLN2 disease, juvenile
 CLN2 disease, infantile
TPPIIntron 7c.887-10A>Gp.Gly296delinsGluAsnProGlyCLN2 disease, juvenile
TPP1Exon 11c.1340G>Ap.Arg447HisCLN2 disease, juvenile
TPPIExon 12c.1439T>Gp.Val480GlyCLN2 disease, juvenile
TPPIExon 12c.1442T>Gp.Phe481CysCLN2 disease, infantile
CLN3Exon 5c.302T>Cp.Leu101ProCLN3 disease, protracted
CLN3Exon 7c.509T>Cp.Leu170ProCLN3 disease, protracted
CLN3Intron 7c.533+1G>CSplice defect/FrameshiftCLN3 disease, protracted
CLN3Exon 8c.597C>Ap.Tyr199XCLN3 disease, protracted
CLN3Intron 9–Intron 13c.791-802_1056+1445del2815p.Gly264_Leu437delinsAlaSerAspSerProAlaSerAlaSerArgValAlaGlyThrThrGly or 2.8 kb deletionCLN3 disease, protracted
CLN3Intron 9–Intron 13Deletion breakpoints not definedp.Gly264_Gln352delinsValCLN3 disease, protracted
 fsX29 
CLN3Exon 11c.883G>Ap.Glu295LysCLN3 disease, protracted
CLN3Exon 13c.1001G>Ap.Arg334HisCLN3 disease, protracted
CLN5Exon 1c.225G>Ap.Trp75XCLN5 disease, juvenile
CLN5Exon 1c.291dupCp.Ser98LeufsX13CLN5 disease, infantile
CLN5Exon 2c.335G>Ap.Arg112HisCLN5 disease, juvenile
CLN5Exon 2c.377G>Ap.Cys126TyrCLN5 disease, adult
CLN5Exon 3c.527_528insAp.Gly177TrpfsX10CLN5 disease, juvenile
CLN5Exon 3c.575A>Gp.Asn192SerCLN5 disease, juvenile
CLN5Exon 3c.620G>Cp.Trp207SerCLN5 disease, juvenile
CLN5Exon 3c.669dupCp.Trp224LeufsX30CLN5 disease, juvenile
CLN5Exon 3c.671G>Ap.Trp224XCLN5 disease, juvenile
CLN5Exon 4c.772T>Gp.Tyr258AspCLN5 disease, juvenile
CLN5Exon 4c.907_1094del188p.Thr303CysfsX10CLN5 disease, juvenile
 CLN5 disease, adult
CLN5Exon 4c.919delAp.Arg307GlufsX29CLN5 disease, juvenile
CLN5Exon 4c.1071_1072delCTp.Leu358AlafsX4CLN5 disease, juvenile
CLN5Exon 4c.1103_1106delAACAp.Lys368SerfsX15CLN5 disease, juvenile
CLN5Exon 4c.1083delTp.Phe361LeufsX4CLN5 disease, juvenile
CLN5Exon 4c.1121A>Gp.Tyr374CysCLN5 disease, adult
CLN6Exon 1c.17G>Cp.Arg6ThrCLN6 disease, Kufs type A
CLN6Exon 2c.139C>Tp.Leu47PheCLN6 disease, Kufs type A
CLN6Exon 2c.150C>Gp.Tyr50XCLN6 disease, Kufs type A
CLN6Exon 3c.200T>Cp.Leu67ProCLN6 disease, Kufs type A
CLN6Exon 3c.231C>Gp.Asn77LysCLN6 disease, Kufs type A
CLN6Exon 3c.308G>Ap.Arg103GlnCLN6 disease, Kufs type A
CLN6Exon 4c.446G>Ap.Arg149HisCLN6 disease, Kufs type A
CLN6Exon 6c.662A>Cp.Tyr221SerCLN6 disease, late infantile variant, protracted
CLN6Exon 6c.662A>Gp.Tyr221CysCLN6 disease, late infantile variant, protracted
CLN6Exon 7c.712_713delinsACp.Phe238ThrCLN6 disease, Kufs type A
CLN6Exon 7c.890delCp.Pro297LeufsX53CLN6 disease, Kufs type A
CLN7Exon 6c.468_469delinsCCp.Ala157ProCLN7 disease, juvenile
CLN8Exon 2c.70C>Gp.Arg24GlyCLN8 disease, EPMR
CLN8Exon 2c.88delGp.Ala30LeufsX20CLN8 disease, late infantile variant, severe
CLN8Intron 2-Exon 3c.544-2566_590p.Ala182AspfsX49CLN8 disease, late infantile variant, severe
  del2613  
CLN8Exon 3c.709G>Ap.Gly237ArgCLN8 disease, EPMR, mild
CTSDExon 5c.685T>Ap.Phe229IleCLN10 disease, juvenile

The majority of CLN1 disease patients, with onset later than infancy, are compound heterozygotes for a typically “severe” mutation and a “mild” mutation. In cases where a patient is homozygote for two “mild” mutations, the phenotype can be even more protracted. For example, two siblings homozygous for p.Thr75Pro with ages at onset at 7 and 9 years, respectively, had a very mild clinical course [Das et al., 1998]. Therefore, the severity of CLN1 disease seems to be dependent on the combination and type of mutations present [Munroe et al., 1998]. Nevertheless, the three CLN1 disease, adult cases are compound heterozygous for a “severe” and a “mild” mutation, rather than being homozygous for two “mild” mutations. The latter could be explained by the hypothesis that other modifiers of phenotype rather than the mutation combination per se are responsible for the final clinical outcome.

CLN2

CLN2 Disease, Clinical Spectrum

Mutations in TPP1 cause three different phenotypes (Table 1). The CLN2 disease, classic late infantile (MIM# 204500), has onset between 2 and 4 years of age, with epilepsy being the leading symptom in most cases [Santavuori, 1988]. Following the initial symptom(s), mental and speech deterioration, myoclonic jerks, ataxia, and visual failure develop [Williams et al., 1999]. Death occurs typically between 6 and 15 years of age [Williams et al., 1999].

The second phenotype, CLN2 disease, juvenile, with disease onset between 6 and 10 years has been described in eight children [Bessa et al., 2008; Elleder et al., 2008; Hartikainen et al., 1999; Kohan et al., 2009; Sleat et al., 1999]. Finally, three cases with CLN2 disease, infantile, and onset below the first year have been reported [Ju et al., 2002; Simonati et al., 2000].

The main ultrastructural finding in CLN2 disease, late infantile, and juvenile is curvilinear bodies (CL), although in approximately 10% of affected children a mixture of CL and fingerprint profiles (FP) is observed [Bessa et al., 2008; Elleder et al., 2008; Hartikainen et al., 1999; Kohan et al., 2009; Palmer et al., 1992; Sleat et al., 1997, 1999]. In all three CLN2 disease, infantile onset cases, there was only CL morphology of the inclusion bodies [Ju et al., 2002; Simonati et al., 2000].

TPP1 Gene and Protein

CLN2 (NM_000391.3) maps to chromosome 11p15 and encodes a lysosomal pepstatin-insensitive carboxypeptidase, tripeptidyl-peptidase I (TPP1) [Sharp et al., 1997; Sleat et al., 1997; Steinfeld et al., 2004]. A precursor form is processed to a 46 kDa mature lysosomal protein of 563 amino acid residues [Sleat et al., 1997]. The protein has a predicted 16 amino acid signal sequence, and contains five potential N-linked glycosylation sites [Sleat et al., 1997; Steinfeld et al., 2002]. The catalytic triad of human TPP1 is Glu272-Asp276-Ser495, with Asp360 being a crucial residue [Mole et al., 2005; Oyama et al., 2005].

TPP1 cleaves tripeptides from the amino terminus of small polypeptides undergoing degradation in the lysosomes, at an optimal pH of 4–4.5 [Vines and Warburton, 1998; Warburton and Bernardini, 2000], but may also possess endopeptidase activity at pH 3 [Ezaki et al., 2000]. Loss of TPP1 activity leads to significant accumulation of subunit c of ATP synthase [Ezaki et al., 2000; Sleat et al., 1997]. Lysosomal storage of subunit c also occurs in other NCL subtypes (though not to the same extent), so this may not be the primary metabolic error in TPP1 deficiency [Fearnley et al., 1990; Palmer et al., 1992; Zhong et al., 1998a].

TPP1 Mutation Spectrum

To date, 89 mutations (23 novel reported here and 66 previously identified) and 22 polymorphisms are known to affect the sequence of TPPI (Table 2, Fig. 2, and Supp. Tables S3 and S4). These consist of 42 missense, 14 nonsense, 17 splice-site affecting, 11 deletions, four insertions, and one deletion–insertion mutations. The residues Ser62, Arg127, Arg206, Arg339, Cys365, Ser475, and Trp548 are affected by more than one mutation (Fig. 2). Originally, it was hypothesized that no TPP1 mutations exist in the first two exons of the gene possibly because this region is not present in the processed mature enzyme due to cleavage [Golabek et al., 2003; Zhong et al., 2000]. The four mutations that are known today to affect this region, are either affecting the splicing or introduce an additional base resulting in frameshifts [Kousi et al., 2009; NCL Mutation Database www.ucl.ac.uk/ncl/mutation; this report]. This suggests that any mutations occurring within this region of TPP1 are more likely to affect processing or stability of the protein [Mole et al., 2005]. The majority of TPP1 mutations result in loss of enzyme activity. Further functional assessment of four missense mutations revealed that these resulted in localization defects (p.Asn286Ser, p.Ile287Asn, p.Thr353Pro, and p.Gln422His), blocking processing to the mature size peptidase and leading to retention of the mutant proteins in the ER where they are rapidly degraded [Steinfeld et al., 2004].

Figure 2.

Schematic representation of the TPP1 gene and the TPP1 protein showing the relative position of the mutations. The coding exons of TPP1 and the TPP1 protein are shown in scale as purple boxes. The untranslated exonic fragments and the introns are shown in gray boxes and lines, respectively, and are not in scale. The missense mutations are shown in the upper and other mutations in the lower part of the gene and the protein. The residues that participate in the formation of the catalytic triad of PPT1 are depicted as green circles. The residues at which N-glycosylation occurs are shown as red circles. The signal peptide is represented with a yellow square. The residues at which the three disulphide bonds occur are marked with yellow text (S1, S2, and S3 are the three disulphide bonds, respectively). The mutations previously reported in the literature are given in black color. The novel mutations reported here are shown in red and the mutations that have only been reported in the NCL database are shown in blue. Nucleotide numbering is based on GenBank reference sequence NM_000391.3.

Among the 89 TPP1 mutations, two common mutations and two genuine founder effects can be distinguished [Ju et al., 2002; Moore et al., 2008; Sleat et al., 1999]. More specifically, the mutations c.509-1G>C and p.Arg208X, which are found worldwide, account for 57% of all identified TPP1 mutant alleles. They are found in 89% of CLN2 disease cases, mostly in compound heterozygosity with other mutations [Sleat et al., 1999; Zhong et al., 1998a]. The p.Gly284Val mutation is established due to a founder effect in the Canadian population, where it accounts for 55% of the CLN2 disease families and 32% of the mutant alleles [Ju et al., 2002], as well as in Newfoundland. This mutation is absent from all the other populations screened, except one patient from the United States who is heterozygous for it [Zhong et al., 2000]. This suggests that p.Gly284Val represents a localized founder effect in both Canada and Newfoundland that most probably occurred in a settler [Moore et al., 2008; Ju et al., 2002].

Nearly half (45%) of the reported CLN2 disease patients are homozygous for their disease-causing mutation, while the remaining 55% are compound heterozygous for different mutations.

Genotype–Phenotype Correlations

Most of the TPP1 mutations cause CLN2 disease, classic late infantile. The missense changes p.Arg127Gln, p.Arg447His, and p.Val480Gly have been associated with more protracted disease phenotypes and later onset, including onset in the juvenile age range [Table 3; Elleder et al., 2008; Sleat et al., 1999; Steinfeld et al., 2002]. The mild disease course associated with p.Arg127Gln is probably due to altered processing of the CLN2 protein [Steinfeld et al., 2002]. The mild clinical phenotype associated with p.Arg447His has been postulated to be due to an unidentified genetic modifier responsible for the variant phenotype [Lin and Lobel, 2001]. From the spectrum of mutations causing CLN2 disease, infantile, the missense p.Phe481Cys could be associated with the earlier onset phenotype [Ju et al., 2002]. The remaining mutations identified in patients with CLN2 disease, infantile (c.509-1G>C, p.Gly284Val, p.Arg208X), are known to cause CLN2 disease, classic late infantile, and thus cannot be unequivocally associated with the infantile phenotype, suggesting that additional genetic factors are likely to contribute to the earlier disease onset and more severe disease course seen in these patients.

CLN3

CLN3 Disease, Clinical Spectrum

CLN3 disease, classic juvenile (MIM# 204200), typically begins with progressive visual loss between 5 and 10 years of age. Around the age of 10–12 years, children start to manifest loss of motor co-ordination, mental decline, and seizures may occur. Later, behavioral abnormalities can develop. In some cases, disease symptoms can be accompanied by hallucinations and/or other neuropsychiatric symptoms. Death usually occurs in the third decade of life [Järvelä et al., 1997]. Retinal examination shows early macular alteration with a fine mottling of the macula or a “bull's eye” with a brownish color of the macula. Later a narrowing of the vessels, bone spicular pigmentations in the peripheral retina, and a pale optic disc appear [Bensaoula et al., 2000; Eksandh et al., 2000]. Diagnosis of CLN3 disease, “classic” juvenile is usually indicated by the presence of vacuolated lymphocytes in peripheral blood [Santavuori, 1988].

Variant forms of CLN3 disease, associated with a slower disease progression, have also been recognized [Table 1; Lauronen et al., 1999; Munroe et al., 1997a; Sarpong et al., 2009; Wisniewski et al., 1998b; Zhong et al., 1998b]. In patients with CLN3 disease, protracted, there may be a long silent period following initial visual failure, free of additional neurological deficits, even for several decades [Åberg et al., 2009; Wisniewski et al., 1998b]. Finally, one case with CLN3 disease, infantile, has been described, with symptoms starting at 5 months of age [de los Reyes et al., 2004].

The intracellular accumulation of autofluorescent storage material in CLN3 disease, classic juvenile appear as typical FP or a combination of FP and CL on EM [Santavuori, 1988]. The major protein component of these abnormal deposits is subunit c of mitochondrial ATPase [Palmer et al., 1992]. Only one CLN3 disease, protracted case, thought to represent a rare variant, with onset at 6–8 years showed GROD morphology of the storage material [Hofman and Taschner, 1995]. No inclusions were identified in muscle and skin biopsy from the patient with CLN3 disease, infantile [de los Reyes et al., 2004].

CLN3 Gene and Protein

In the CLN3 gene (NM_000086.2) that maps to chromosome 16, a core risk haplotype was identified on 73% of disease chromosomes associated with the common 1-kb deletion [Eiberg et al., 1989; The International Batten Disease Consortium, 1995]. CLN3 encodes a predicted 438 amino acid protein product with a molecular mass of 48 kDa, and does not bear significant similarities to proteins of known function [The International Batten Disease Consortium, 1995].

CLN3 is a transmembrane protein composed of six membrane-spanning domains, having both N- and C-terminal tails facing the cytoplasm [Ezaki et al., 2003; Kyttälä et al., 2004; Phillips et al., 2005] and an amphipathic helix facing the lysosomal/endosomal lumen [Nugent et al., 2008]. CLN3 resides in the endosomes/lysosomes of nonneuronal cells, and in the synaptic vesicles of neuronal cells [Haskell et al., 1999]. Targeting studies demonstrated the existence of two lysosomal sorting signals in the second cytoplasmic loop domain and the C-terminal tail of CLN3, respectively [Kyttälä et al., 2004, 2005; Storch et al., 2004, 2007].

CLN3 Mutation Spectrum

Altogether 59 disease-associated mutations (13 novel reported here and 46 previously reported) and nine polymorphisms are known to occur in CLN3 (Table 2, Fig. 3, and Supp. Tables S5 and S6). These include 12 missense, 13 nonsense, 16 splice-site affecting, 11 deletions, six insertions, and one mutation affecting the first methionine. Approximately 80% of the CLN3 mutations produce prematurely truncated products. The relatively large number of insertion and deletion mutations may be due to an overrepresentation of Alu elements in CLN3 (one element every 0.7 kb over the CLN3 region vs. one element every 5 kb across the human genome) [Mitchison et al., 1997; Munroe et al., 1997b]. The CLN3 missense mutations are believed to exert their effects by impairing protein function, since none of the mutations studied (p.Leu101Pro, p.Leu170Pro, p.Glu295Lys, p.Val330Phe, and p.Arg334His) interfered with the normal targeting of CLN3 to the lysosomes [Haskell et al., 2000]. Most single amino acid mutations that are not located within membrane-spanning domains reside on the topological face of CLN3 that is exposed to the lysosomal/endosomal lumen, suggesting that this face of the protein is functionally important [Fig. 3; Nugent et al., 2008]. Two distinct mutations occur in each of the residues Gly187, Glu295, and Arg334.

Figure 3.

Schematic representation of the CLN3 gene and the CLN3 protein showing the relative position of the mutations. The coding exons of CLN3 are shown in scale as purple boxes, while the untranslated exonic fragments and the introns are shown in gray boxes and lines, respectively, and are not in scale. The transmembrane domains of CLN3 are shown as purple boxes and the cytosolic and lysosomal facing domains as lines and are in scale. The lysosomal membrane is depicted as an orange rectangle in which CLN3 is anchored. The residues carrying lysosomal targeting information are given in the yellow squares. The missense mutations are shown in the upper and other mutations in the lower part of the gene. The mutations previously reported in the literature are given in black color. The novel mutations reported here are shown in red and the mutations that have only been reported in the NCL database are shown in blue. Nucleotide numbering is based on GenBank reference sequence NM_000086.2.

The c.461-280_677+382del966 (1-kb deletion) mutation accounts for 81% of known disease chromosomes [The International Batten Disease Consortium, 1995]. The 1-kb deletion results in a predicted truncated protein of 181 amino acids or a spliced variant that lacks exons 7–8 [Kitzmuller et al., 2008; The International Batten Disease Consortium, 1995]. Its association almost always with a common haplotype suggests that it derived from a common founder and occurred over 2,000 years ago [Mitchison et al., 1995]. Studies showed that this mutant protein retains some of its function [Kitzmuller et al., 2008] but is unable to traffic out of the ER [Järvelä et al., 1999], possibly due to loss of two of the identified lysosomal sorting signals.

Genotype–Phenotype Correlations

Patients homozygous for the 1-kb deletion manifest CLN3 disease, classic juvenile [Lauronen et al., 1999; Munroe et al., 1997a]. Although initially reported that compound heterozygous patients for the 1-kb deletion display a potentially slower disease course [Järvelä et al., 1997; Lauronen et al., 1999], a recent study evaluating phenotypic differences among patients that were homozygous, compound heterozygous for the 1-kb deletion or homozygous for p.Arg334His, did not find any differences between the three groups that would suggest possible genotype–phenotype correlations [Adams et al., 2010]. In the light of these findings, the delayed classic CLN3 disease phenotype displayed by compound heterozygous patients for the 1-kb deletion and the 2.8-kb deletion, removing exons 9 through 13, is more likely to be due to the 2.8-kb deletion [Munroe et al., 1997a].

Some missense mutations are regarded as “mild” mutations, since they allow retention of significant levels of function of mutant proteins in yeast [Table 3; Haines et al., 2009]. Accordingly, compound heterozygotes for the 1-kb deletion and some of the reported missense mutations (p.Leu101Pro, p.Leu170Pro, and p.Glu295Lys) present with protracted CLN3 disease [Lauronen et al., 1999; Munroe et al., 1997a; Zhong et al., 1998b]. In contrast, mutations such as p.Gly187Ala and p.Val330Phe in some patients are reported to be associated with CLN3 disease, classic juvenile, when found in compound heterozygosity with the 1-kb deletion [Haskell et al., 2000; Munroe et al., 1997a]. The missense p.Arg334His can be associated with both CLN3 disease, classic juvenile, as well as CLN3 disease, protracted [Munroe et al., 1997a].

Although the molecular diagnosis of a patient reported to have CLN3 disease, infantile is not complete, with only one mutation in CLN3 having been identified (1-kb deletion), the patient was classified with CLN3 disease because of the unequivocal association of the 1-kb deletion with CLN3 disease, juvenile [de los Reyes et al., 2004]. It can be argued that a second yet unidentified mutation in CLN3, and/or additional genetic determinants, are responsible for the more aggressive disease course and earlier onset of symptoms described in this case.

The nonsense mutation p.Tyr199X, identified in a Lebanese family with five affected siblings, is associated with a milder CLN3 disease, protracted phenotype [Sarpong et al., 2009]. A possible explanation for this milder phenotype associated with p.Tyr199X mutant protein as opposed to the 1-kb deletion mutant protein, could be that the former contains at least 60 amino acid residues from the structurally and functionally important region encoded by exons 7–8, which are missing from the 1-kb deletion mutant proteins [Sarpong et al., 2009].

CLN5

CLN5 Disease, Clinical Spectrum

The CLN5 disease, late infantile variant phenotype (MIM# 256731), was first described in children from Finland who share the same founder mutation. The onset is at 4–7 years of age, with motor clumsiness as the most common presenting symptom [Santavuori et al., 1982, 1993]. As the disease progresses, epileptic seizures, mental and motor deterioration, visual impairment, ataxia, and myoclonus develop [Holmberg et al., 2000; Santavuori et al., 1982, 1991]. Affected children become blind between 7 and 10 years and death usually occurs between 10 and 30 years, although one patient survived until the age of 41 [Holmberg et al., 2000; Moore et al., 2008]. In brain imaging studies, cerebellar atrophy is pronounced [Holmberg et al., 2000].

Three atypical phenotypes can be recognized (Table 1). CLN5 disease, juvenile, has an onset between 4 and 9 years with visual failure, loss of strength, and tremor of lower limbs [Cannelli et al., 2007; Kohan et al., 2008; Lebrun et al., 2009; Pineda-Trujillo et al., 2005; Xin et al., 2010]. The patients develop the full spectrum of symptoms within 1 to 8 years from initial clinical presentation [Xin et al., 2010]. In the CLN5 disease, adult phenotype described in two cases the onset was delayed until age 17 years and death occurred in the fourth decade of life [Sleat et al., 2009; Xin et al., 2010]. Finally, one case of CLN5 disease, infantile with onset at 4 months, has been reported [Cismondi et al., 2008].

CLN5 disease, late infantile variant, and juvenile, are typically associated with FP morphology of the storage granules [Pineda-Trujillo et al., 2005; Santavuori et al., 1982]. Occasionally FP profiles are admixed with CL or rectilinear bodies (RL) [Mole et al., 2005]. CLN5 disease, infantile, and adult subtypes, are associated with a GROD lipopigment morphology admixed with FP [Cismondi et al., 2008; Xin et al., 2010]. Subunit c of the mitochondrial ATP synthase is the main protein component of the storage material in CLN5 disease [Tyynelä et al., 1997].

CLN5 Gene and Protein

In CLN5 (NM_006493.2) that maps to chromosome 13q21.1-q32, translation can occur from any of four initiator methionines at positions p.1, p.30, p.50, and p.62, producing CLN5 polypeptides with expected molecular masses of 40.3, 41.5, 43.4, and 46.3, respectively [Isosomppi et al., 2002; Savukoski et al., 1998]. When translation starts from p.Met1, the peptide produced is composed of 407 amino acids, having a signal peptide at the N-terminus that is cleaved at position p.96 [Schmiedt et al., 2010]. CLN5 shares no homology with known genes or proteins [Isosomppi et al., 2002; Savukoski et al., 1998].

It is now generally accepted that CLN5 is a soluble lysosomal glycoprotein [Holmberg et al., 2004; Isosomppi et al., 2002; Schmiedt et al., 2010]. Lysosomal sorting of CLN5 occurs mainly via the Man6P dependent pathway, although alternative sorting routes have also been proposed [Kollmann et al., 2005; Schmiedt et al., 2010; Sleat et al., 2009].

CLN5 Mutation Spectrum

To date, 33 disease-associated mutations (nine novel and 24 previously reported) and eight polymorphisms are known to occur in the sequence of CLN5 (Table 2, Fig. 4, and Supp. Table S7 and S8). These comprise 15 missense, eight nonsense, seven deletion, and three insertion mutations. Approximately 53% of the pathogenic mutations result in prematurely terminated transcripts that cause mRNA instability, folding defects, degradation via nonsense-mediated decay, reduced protein expression, or impaired sorting of transcripts that get translated [Bessa et al., 2006; Lebrun et al., 2009; Xin et al., 2010]. The missense mutations that have been studied (p.Arg112His, p.Arg112Pro, p.Asp279Asn and p.Trp379Cys) have been reported to cause ER retention of the mutant polypeptides, perhaps due to misfolding [Lebrun et al., 2009; Schmiedt et al., 2010]. Each of residues Trp75, Arg112, Trp207, and Trp224 is affected by two different mutations (Fig. 4).

Figure 4.

Schematic representation of the CLN5 gene and the CLN5 protein showing the relative position of the mutations. The coding exons of CLN5 and the CLN5 protein are shown in scale as purple boxes. The untranslated exonic fragments and the introns are shown in gray boxes and lines, respectively, and are not in scale. The signal peptide is represented with a yellow square. The missense mutations are shown in the upper and other mutations in the lower part of the gene and the protein. The mutations previously reported in the literature are given in black color. The novel mutations reported here are shown in red. Nucleotide numbering is based on GenBank reference sequence NM_006493.2.

The nonsense mutation p.Tyr392X is a founder mutation present in 94% (34/36) of Finnish disease chromosomes [Savukoski et al., 1998]. Although mutations in CLN5 were initially reported to be confined to Finnish patients (Finnish variant late infantile NCL; Table 1), the expansion of the mutation spectrum of CLN5 disease in recent years by identification of defects in patients from 15 different countries emphasizes the worldwide distribution of NCL defects (Supp. Table S7 and S8).

Genotype–Phenotype Correlations

In CLN5 disease there is little variation in the clinical phenotype, brain imaging data, neurophysiologic, and/or histopathologic findings of patients carrying different CLN5 mutations [Holmberg et al., 2000]. Mutation p.Ser98LeufsX13 is associated with infantile onset disease (CLN5 disease, infantile). Patients with CLN5 disease, juvenile, can carry an apparently deleterious mutation (nonsense, frameshift causing) on both alleles. Mutation p.Tyr374Cys identified in homozygosity in two patients with CLN5 disease, adult, seems to have a “mild” impact on the protein, probably allowing for some residual CLN5 function [Xin et al., 2010]. One of these individuals also carries a single change in CLCN6, which may also contribute in modification of the disease phenotype [Poet et al., 2006].

CLN6

CLN6 Disease, Clinical Spectrum

In CLN6 disease, late infantile variant (MIM# 601780), the onset is slightly later than in CLN2 disease, typically between 3 and 8 years of age [Table 1; Mole et al., 2005]. Seizures and motor difficulties present early, followed by myoclonus, speech impairment, ataxia, and mental regression [Mole et al., 2005]. Visual impairment becomes apparent later than in patients with CLN2 disease [Moore et al., 2008]. Patients lose vision and all motor skills between 4 and 10 years [Teixeira et al., 2003a]. Death occurs by the mid 20s [Pena et al., 2001].

Very recently, mutations in CLN6 have been associated with adult onset Kufs disease type A [Arsov et al., 2011]. Patients with CLN6 disease, adult, present around the age of 30 years (range 16–51 years) with progressive myoclonus epilepsy, in contrast to those diagnosed with Kufs disease type B who present with dementia and motor dysfunction [Arsov et al., 2011]. Both recessive and dominant forms of Kufs type A have been described, however mutations in CLN6 cause only recessive disease [Arsov et al., 2011].

Although not reported to represent atypical phenotypes patients with CLN6 disease, late infantile and onset ranging from 18 months to 9 years of age have been reported [Elleder et al., 1997a; Teixeira et al., 2003b]. Examples of inter- and intrafamilial variation in CLN6 disease have been reported, suggesting that additional factors influence the clinical manifestation [Cannelli et al., 2009].

Mutations in CLN6 are always associated with an admixed morphology of the storage material consisting of FP and CL, characteristic of the group of variant late infantile subtypes [Sharp et al., 2003; Topcu et al., 2004]. In patients with CLN6 disease, Kufs type A, the storage material identified in pathological material consisted either of FP or GRODs [Arsov et al., 2011]. The main accumulating protein component is mitochondrial ATPase subunit c [Elleder et al., 1997b; Mole et al., 2005].

CLN6 Gene and Protein

Mutations in CLN6 that maps to chromosome 15q21-q23, were originally identified in Costa Rican and Venezuelan patients [NM_017882.2; Gao et al., 2002; Sharp et al., 1997; Wheeler et al., 2002]. A single-base pair insertion has also been detected in the nclf mouse, a naturally occurring mouse model for CLN6 [Gao et al., 2002; Wheeler et al., 2002]. CLN6 encodes a predicted 311 amino acid transmembrane protein and shares no homology with known proteins or functional domains [Gao et al., 2002; Sharp et al., 2003; Wheeler et al., 2002].

CLN6 consists of seven membrane-spanning domains, with the N-terminus facing the cytoplasm and the C-terminus directed toward the ER lumen [Heine et al., 2007]. In both neuronal and nonneuronal cells CLN6 resides in the ER [Heine et al., 2007; Mole et al., 2004]. The localization of CLN6 is mediated by two unusual ER retention signals, one comprising the first 49 N-terminal amino acids and the second being in the TM6-TM7 domains [Heine et al., 2007; Mole et al., 2004; Teixeira et al., 2006]. Defective CLN6 function is associated with impaired lysosomal acidification [Heine et al., 2004; Holopainen et al., 2001]. Given that both protein and lipid synthesis take place in the ER, it has been hypothesized that CLN6 mediates selective transport of proteins or lipids that are essential for lysosomal function and acidification [Mole et al., 2004].

CLN6 Mutation Spectrum

Altogether 63 mutations (15 novel reported here and 48 previously reported) and seven polymorphisms are known to affect CLN6 (Table 2, Fig. 5, and Supp. Tables S9 and S10). These consist of 38 missense, five nonsense, five splice-site affecting, one deletion–insertion, four insertion, and 10 deletion mutations (one of the deletions having undefined reported breakpoints c.83+?_297+?del). Two distinct mutations affect each of the residues Arg62, Glu72, Asp83, Tyr172, Gly259, and Pro297. The fact that three different mutations affect the p.Tyr221 residue indicates the potential existence of a mutation hotspot. With the exception of seven mutations that affect the cytosolic N-terminal protein domain, all other amino acid changing mutations reside on the topological face of CLN6 that is exposed to the ER lumen (Fig. 5), implying an important functional role for this surface, possibly serving for interaction with other ER proteins [Mole et al., 2004]. Prematurely truncating mutations result in reduction of gene expression [Cannelli et al., 2009]. None of the mutants evaluated prevents correct localization or retention of CLN6 in the ER [Mole et al., 2004], nor dimerization [Kurze et al., 2010]. Nevertheless, the rate of synthesis and stability of these CLN6 mutants are reduced in comparison with wild-type CLN6 [Kurze et al., 2010].

Figure 5.

Schematic representation of the CLN6 gene and the CLN6 protein showing the relative position of the mutations. The coding exons of CLN6 are shown in scale as purple boxes, while the untranslated exonic fragments and the introns are shown in gray boxes and lines, respectively, and are not in scale. The transmembrane domains of CLN6 are shown as purple boxes and the cytosolic and ER lumen facing domains as lines and are in scale. The ER membrane is depicted as an orange rectangle in which CLN6 is anchored. The signal peptide sequences are highlighted with a yellow square. In the gene representation, the missense mutations are shown in the upper and other mutations in the lower part of the gene. The mutations previously reported in the literature are given in black color. The novel mutations reported here are shown in red and the mutations that have only been reported in the NCL database are shown in blue. Nucleotide numbering is based on GenBank reference sequence NM_017882.2.

Three CLN6 mutations are enriched in specific populations. The nonsense mutation p.Glu72X was present in 93% of the disease alleles in Costa Rican patients [Gao et al., 2002; Wheeler et al., 2002]. The p.Ile154del mutation was identified in 94% of Portuguese disease chromosomes [Teixeira et al., 2003a]. Finally, in Newfoundland all CLN6 disease patients are homozygous for p.Val91GlufsX42 [Moore et al., 2008].

Genotype–Phenotype Correlations

The age of onset for CLN6 disease is now known to be very wide. The striking differences in disease severity arising from mutations in this gene can be explained by the existence of mutations that completely abolish CLN6 function, and those that are “milder” resulting in retention of partial function. Approximately 40% of the patients with CLN6 disease, late infantile variant, carry mutations that either predict premature truncation or a severely abnormal protein in both disease alleles [Arsov et al., 2011]. In contrast, in patients with CLN6 disease, Kufs type A, such mutations were only identified in compound heterozygosity with “mild” mutations in two cases [Arsov et al., 2011]. The changes at the tyrosine residue at position 221 (p.Tyr221Ser and p.Tyr221Cys) might also represent “mild” mutations, since they have been associated with a more protracted clinical course that does not involve the development of a visual phenotype [Table 3; Cannelli et al., 2009; Sharp et al., 2003].

The majority of CLN6 disease patients (73%) are homozygous for their mutation. The mutations associated with CLN6 disease, Kufs type A, do not seem to cluster to a specific protein domain that suggests that the functionally important sites of CLN6 remain intact.

CLN7

CLN7 Disease, Clinical Spectrum

The CLN7 disease, late infantile variant phenotype (MIM# 610951), is essentially indistinguishable from that of other late infantile NCL forms (CLN2, CLN5, CLN6, or CLN8). Disease onset is between the age of 2 and 7 years. The most common initial symptoms are seizures and developmental regression [Kousi et al., 2009; Topcu et al., 2004]. However, patients with ataxia or visual impairment as initial symptoms have also been reported [Topcu et al., 2004]. Disease progression is rapid with mental and motor regression, loss of vision, and seizures developing in the majority of patients early in the disease course [Kousi et al., 2009]. Most patients lose the ability to walk an average of 2 years after disease onset, and all die prematurely [Topcu et al., 2004]. Compared to classical CLN2 disease, CLN7 disease shows a somewhat later onset and a more severe seizure phenotype [Siintola et al., 2007; Topcu et al., 2004], with cerebral and cerebellar atrophy having a slower course.

Only one atypical case, CLN7 disease, juvenile, has thus far been described (Table 1). In this patient, visual failure at the age of 11 years was the initial symptom and the disease course was protracted, the patient being wheelchair bound but still alive at the age of 43 years [Kousi et al., 2009].

CLN7 disease is associated with condensed FP in circulating lymphocytes [Siintola et al., 2007; Topcu et al., 2004]. In addition, a complex of FP profiles and RL inclusions, occasionally associated with CL profiles or GRODs has been reported [Aiello et al., 2009; Kousi et al., 2009].

MFSD8 Gene and Protein

The major facilitator superfamily domain 8-containing gene (MFSD8; NM_152778.2) identified in the CLN7 locus, on chromosome 4q28.1-q28.2, encodes a 518 amino acid protein [Siintola et al., 2007]. MFSD8 has a molecular mass of approximately 58 kDa and possesses two N-glycosylation consensus sites [Sharifi et al., 2010; Siintola et al., 2007; Steenhuis et al., 2010].

Bioinformatic analysis of MFSD8 predicts 12 transmembrane domains with both the N- and C-termini extending into the cytosol [Siintola et al., 2007]. MFSD8 is targeted to the lysosomes via a main dileucine-based motif at the N-terminal portion of the protein [Sharifi et al., 2010; Siintola et al., 2007; Steenhuis et al., 2010]. However, additional and presumably unconventional motifs are thought to partially direct MFSD8 to the lysosome [Sharifi et al., 2010]. Based on sequence homology analyses, MFSD8 belongs to the major facilitator superfamily of transporter proteins and is thus likely to function as a lysosomal transporter [Siintola et al., 2007]. However, the substrate(s) that MFSD8 transports remains to be identified.

CLN7/MFSD8 Mutation Spectrum

Altogether 31 mutations (22 previously reported and nine novel) and three coding polymorphisms are known to affect the sequence of MFSD8 (Table 2, Fig. 6, and Supp. Tables S11 and S12). These involve 15 missense, six nonsense, one frameshift, and six splice-site affecting mutations, one mutation affecting the initiator methionine, as well as two mutations that could be either missense or interfere with the correct splicing of the transcript. The residues Thr160 and Arg465 are affected each by two different missense mutations. Localization studies have shown that none of the missense mutations studied interfere with correct localization of MFSD8 [Kousi et al., 2009; Sharifi et al., 2010; Siintola et al., 2007]. This implies that disturbed functional properties, rather than altered subcellular localization, are the primary consequence of the missense mutations [Kousi et al., 2009].

Figure 6.

Schematic representation of the MFSD8 gene and the MFSD8 protein showing the relative position of the mutations. The coding exons of MFSD8 are shown in scale as purple boxes, while the untranslated exonic fragments and the introns are shown in gray boxes and lines, respectively, and are not in scale. The transmembrane domains of MFSD8 are shown as purple boxes and the cytosolic and lysosomal facing domains as lines and are in scale. The lysosomal membrane is depicted as an orange rectangle in which MFSD8 is anchored. The residues at which N-glycosylation occurs are shown as red circles. In the gene representation, the missense mutations are shown in the upper and other mutations in the lower part of the gene. The mutations previously reported in the literature are given in black color. The novel mutations reported here are shown in red. The major lysosomal targeting motif is highlighted with a yellow square. Nucleotide numbering is based on GenBank reference sequence NM_152778.2.

Approximately 94% of the MFSD8 mutations are private [Aiello et al., 2009; Aldahmesh et al., 2009; Kousi et al., 2009; Siintola et al., 2007; Stogmann et al., 2009]. A founder effect has been detected in a group of Roma patients originating from the former Czechoslovakia [Kousi et al., 2009]. Fourteen of the 15 Roma patients were homozygous for the missense mutation p.Lys294Thr and shared an identical haplotype on which the mutation occurred [Kousi et al., 2009]. Finally, the splice-site affecting mutation c.863+3_4insT occurred in three apparently unrelated patients from Italy, suggesting a shared ancestry [Aiello et al., 2009].

Genotype–Phenotype Correlations

The MFSD8 mutations are associated with a relatively uniform clinical manifestation compatible with a complete loss of gene function [Kousi et al., 2009]. The only exception is the in-frame deletion–insertion mutation c.468_469delinsCC resulting in a single amino acid substitution (p.Ala157Pro), which is associated with a protracted CLN7 disease, juvenile, phenotype [Table 3; Kousi et al., 2009].

CLN8

CLN8 Disease, Clinical Spectrum

CLN8 disease (MIM# 600143) is associated with two distinct phenotypes: CLN8 disease, EPMR (progressive epilepsy with mental retardation), first identified in Finland [Herva et al., 2000; Hirvasniemi et al., 1994]; and a CLN8 disease, late infantile variant, first described in a subset of Turkish patients [Table 1; Mitchell et al., 2001; Topcu et al., 2004]. The CLN8 disease, late infantile variant, shows an earlier onset and more rapidly progressing disease course than EPMR, with a clinical presentation similar to the other variant late infantile NCL forms [Ranta et al., 2004; Topcu et al., 2004]. The onset is around 2–7 years of age. Patients show a rapid disease progression involving ataxia, myoclonus, speech delay, developmental regression, visual failure, and seizures within 2 years of onset and loss of ambulation [Ranta et al., 2004; Topcu et al., 2004].

CLN8 disease, EPMR, is characterized by normal early development, onset of drug resistant generalized tonic–clonic seizures between the ages of 5 and 10 years, and subsequent mental retardation. The seizures increase in frequency until puberty from when the epileptic activity starts to decline, although does not remit completely. Mental retardation begins 2–5 years after the onset of seizures. The decline in the intellectual level is most rapid before adulthood, leading to mental retardation by the age of 30 at the latest [Hirvasniemi et al., 1994]. Visual loss is not a prominent feature of CLN8 disease, EPMR. Death occurs after the fifth decade of life [Ranta and Lehesjoki, 2000].

CLN8 disease, is associated with condensed FP or CL-like storage morphology, a complex of both CL and FP, and occasionally with RL and/or GRODs [Herva et al., 2000; Topcu et al., 2004]. The main protein component accumulating is mitochondrial ATP synthase subunit c [Herva et al., 2000].

CLN8 Gene and Protein

CLN8 (NM_018941.3) maps to 8p23 and was first identified as the gene underlying EPMR [Ranta et al., 1999; Tahvanainen et al., 1994]. A homozygous 1-bp insertion mutation (c.267_268insC) has also been identified in the motor neuron degeneration mouse (mnd) that is a naturally occurring mouse model for CLN8 disease [Ranta et al., 1999]. CLN8 encodes a putative membrane protein of 286 amino acids, with five predicted transmembrane domains and unknown function [Lonka et al., 2000; Ranta et al., 1999].

CLN8 is a member of the TLC (TRAM-LAG1-CLN8) protein family that is thought to have a role in sensing, biosynthesis, and metabolism of lipids or protection of proteins from proteolysis [Winter and Ponting, 2002]. The TLC domain in CLN8 is composed of the residues p.62-262 [Winter and Ponting, 2002]. CLN8 is a resident of the ER and the ER–Golgi intermediate compartment (ERGIC) in nonneuronal cells, while a location outside the ER has been suggested in neurons [Lonka et al., 2004]. A C-terminal ER-retrieval signal (283-KKPR-286) is responsible for the intracellular localization of CLN8 and for its ability to recycle between the ER and ERGIC compartments [Lonka et al., 2000].

CLN8 Mutation Spectrum

Altogether 25 mutations (16 previously reported and nine novel) are known to affect the sequence of CLN8. These involve 20 missense and five deletion mutations (Table 2, Fig. 7, and Supp. Tables S13 and S14). The amino acid residues Ala30 and Arg204 are affected each by two different mutations (Fig. 7). None of the six missense mutations studied (p.Arg24Gly, p.Ala30Pro, p.Tyr158Cys p.Gln194Arg, p.Arg204Cys, and p.Trp263Cys) affected correct protein targeting, suggesting that they disturb the functional properties of CLN8 [Lonka et al., 2000, 2004; Vantaggiato et al., 2009]. The genomic deletion of 2,566 nucleotides is predicted to result in mRNA instability, since no mutant transcripts could be detected in cells from the patient [Reinhardt et al., 2010].

Figure 7.

Schematic representation of the CLN8 gene and the CLN8 protein showing the relative position of the mutations. The coding exons of CLN8 are shown in scale as purple boxes, while the untranslated exonic fragments and the introns are shown in gray boxes and black lines, respectively, and are not in scale. The transmembrane domains of CLN8 are shown as purple boxes and the cytosolic and ER lumenal facing domains as lines and are in scale. The ER membrane is depicted as an orange rectangle in which CLN8 is anchored. The TLC domain spans the region from p.62-p.262. The ER retrieval signal is shown in the yellow square. In the gene representation, the missense mutations are shown in the upper and other mutations in the lower part of the gene. The mutations previously reported in the literature are given in black color. The novel mutations reported here are shown in red. Nucleotide numbering is based on GenBank reference sequence NM_018941.3.

The missense p.Arg24Gly that causes CLN8 disease, EPMR, in Finnish patients represents a founder mutation [Ranta et al., 1999]. The majority of mutations affecting CLN8 are private.

Genotype–Phenotype Correlations

Homozygosity for mutation p.Arg24Gly is associated with a protracted clinical course of EPMR that is not associated with myoclonus or visual failure [Table 3, Ranta et al., 2004]. In a single Finnish patient who was compound heterozygous for p.Arg24Gly and p.Gly237Arg, the clinical phenotype was reported to be even more protracted than CLN8 disease, EPMR [Siintola et al., 2006a]. All the other CLN8 mutations are associated with the more severe CLN8 disease, late infantile variant phenotype. Mutations c.88delG and c.544-2566_590del are reported to be associated with an earlier onset and more progressive disease course, but when these or similar mutations (like c.66delG) are found in compound heterozygosity with nontruncating changes, a slightly milder phenotype is observed [Reinhardt et al., 2010]. Although initially hypothesized that mutations affecting the TLC domain might cause more aggressive phenotypes, it is now accepted that mutations in different protein domains do not relate to differences in clinical severity.

CLN10

CLN10 Disease, Clinical Spectrum

Only 10 cases of CLN10 disease, congenital (MIM# 610127), have been reported in the literature [Barohn et al., 1992; Brown et al., 1954; Fritchie et al., 2009; Garborg et al., 1987; Norman and Wood, 1941; Sandbank, 1968; Siintola et al., 2006b]. Patients present with postnatal respiratory insufficiency and status epilepticus while seizures probably occur already before birth. Death occurs within hours to weeks after birth. All patients have microcephaly and on postmortem examination the patient brains appear small and firm with the gyral pattern, suggesting developmental delay already from the 30th to 32nd week of gestation [Fritchie et al., 2009; Sandbank, 1968].

In a single patient a significantly milder condition has been described, resembling juvenile-onset NCL and termed CLN10 disease, juvenile [Table 1; Steinfeld et al., 2006]. In this patient disease onset was at early school age, the patient being still alive at the age of 17 years. The presenting symptoms included ataxia and visual disturbances. The patient progressively developed cognitive decline, loss of speech, vision, and motor functions [Steinfeld et al., 2006].

CTSD mutations are always associated with GROD morphology [Barohn et al., 1992; Fritchie et al., 2009; Garborg et al., 1987; Humphreys et al., 1985; Steinfeld et al., 2006].

CTSD Gene and Protein

The cathepsin D gene (CTSD) was first found to underlie a form of NCL in a naturally occurring ovine animal model for congenital NCL [Tyynelä et al., 2000], with mutations subsequently described in human patients [NM_001909.4; Fritchie et al., 2009; Siintola et al., 2006b; Steinfeld et al., 2006].

CTSD belongs to the pepsin family of proteases [Metcalf and Fusek, 1993]. The final mature protein is comprised of two polypeptides, the heavy and light chains containing 196 and 141 amino acid residues, respectively [Faust et al., 1985; Rawlings and Barrett, 1995]. Mature CTSD functions as an aspartyl protease that cleaves peptide bonds flanked by bulky hydrophobic amino acids inside the polypeptide chains. Its function consists in mediating protein degradation, protease precursor form activation, or protease inhibitor inactivation [Scarborough and Dunn, 1994]. The enzyme has a bilobed structure with the active site cleft located between the lobes [Metcalf and Fusek, 1993]. The catalytic site of CTSD is comprised of two aspartic acid residues (Asp97 and Asp294), one on each chain of CTSD [Hunt and Dayhoff, 1970; Metcalf and Fusek, 1993].

Lysosomal targeting of CTSD occurs via the Man6P-receptor-dependent pathway, although Man6P-receptor-independent mechanisms have also been proposed to participate in the process [Zaidi et al., 2008].

CTSD Mutation Spectrum

Only four mutations, three missense and one nonsense, have been reported to date [Fig. 8 and Supp. Tables S15 and S16; Fritchie et al., 2009; Siintola et al., 2006b; Steinfeld et al., 2006]. All are associated with reduced CTSD enzymatic activity. With the exception of p.Trp383Cys, all mutant polypeptides are reported to be normally processed to the mature protein. The p.Thr383Cys mutant proteins have been shown to be mistargeted to the ER compartment, while p.Phe229Ile mutants were not found to interfere with correct protein localization [Fritchie et al., 2009; Steinfeld et al., 2006]. No common mutations are yet recognized in CLN10 disease.

Figure 8.

Schematic representation of the CTSD gene and the CTSD protein showing the relative position of the mutations. The coding exons of CTSD and the CTSD protein are shown in scale as numbered boxes. The untranslated exonic fragments and the introns are shown in nonnumbered boxes and black lines, respectively, and are not in scale. The residues that participate in the formation of the catalytic site of CTSD are depicted as circles. The residues at which the four disulphide bonds occur are marked as S1, S2, S3, and S4, respectively. The lysosomal signal is indicated. In the gene representation, the missense mutations are shown in the upper and other mutations in the lower part of the gene and the protein. All CTSD affecting mutations have previously been reported in the literature. Nucleotide numbering is based on GenBank reference sequence NM_001909.4.

Genotype–Phenotype Correlations

All the truncating and missense mutations that completely abolish the enzymatic activity of CTSD correlate with the severe CLN10 disease, congenital form [Siintola et al., 2006b; Tyynelä et al., 2000]. Only the p.Phe229Ile mutation retains some residual enzyme activity and is thus associated with the milder phenotype CLN10 disease, juvenile [Steinfeld et al., 2006].

Other Types of NCL Disease

Despite recent advances in molecular genetics applied to patients that fulfill the clinical and histopathological criteria of NCL, in certain patients the underlying molecular defect remains unknown. For some with a possibly distinct phenotype, the mutations are assumed to be in genes that, although not yet identified, are referred to colloquially by a CLN name. Consistent with this, identification of CLN4, CLN9, CLN11, and perhaps even beyond are expected in the future.

CLN4 refers to adult onset NCL (MIM# 204300), with either autosomal recessive (Kufs disease) or autosomal dominant (Parry disease) inheritance [Berkovic et al., 1988; Nijssen et al., 2002]. The clinical onset of adult onset NCL is at about 30 years of age, with patients developing progressive dementia, ataxia, seizures, and myoclonus but no ocular involvement. Death occurs prematurely on average 13 years after the disease onset [Sleat et al., 2009]. The storage material can be of GROD, CL, RL, and/or FP morphology, and the accumulating protein can be either SAP A and D or the subunit c of the mitochondrial ATP synthase [Hall et al., 1991; Nijssen et al., 2003]. It is likely that this group of adult onset NCL is caused by mutations in more than one genes.

CLN9 is provisionally assigned to a variant juvenile onset NCL disease subtype (MIM# 609055) [Schulz et al., 2004].

Two late onset patients diagnosed with NCL are reported to carry a single missense mutation in CLCN6 on one chromosome only [Poet et al., 2006]. One of these patients was subsequently found to carry two mutations in CLN5 [Xin et al., 2010] and therefore is now diagnosed with CLN5 disease. The possibility remains that CLCN6 and perhaps other chloride-channel genes act as modifiers of a given NCL phenotype, ameliorating the course of disease in this specific case. This view is supported by the fact that in mice-targeted disruption of Clcn6 results in progressive lysosomal storage associated with accumulation of lipofuscin and subunit c of mitochondrial ATP synthase [Poet et al., 2006]. However, lysosomal pH is normal, and there is no visual impairment, nor loss of neurons [Poet et al., 2006].

One patient diagnosed with adult onset NCL and GROD lipopigment morphology was found to be heterozygous for the missense changes p.Glu355Lys and p.Ser298Pro in the gene SGSH, which encodes the enzyme N-sulphoglucosamine sulphohydrolase, also known as sulphamidase or heparan sulfate sulfatase [Sleat et al., 2009]. Mutations in this gene are typically associated with the childhood lysosomal storage disorder mucopolysaccharidosis type IIIA [MPSIIIA; Scott et al., 1995]. Therefore, other diagnosed NCL cases, perhaps particularly those of later onset, may have mutations in genes more typically causing other related genetic diseases.

NCL Gene Interaction and Interplay

Whether there is a common disease pathway in the group of NCL disorders remains an unanswered question. However, there are interactions between TPP1 and CLN5 [Vesa et al., 2002], and concerted trafficking of PPT1 and CLN5 via a Man6P-independent pathway [Lyly et al., 2009; von Schantz et al., 2008]. CLN5 seems to have a central role because it was shown to interact with most NCL proteins [Lyly et al., 2009; Vesa et al., 2002]. The interactions of CLN5 with PPT1, CLN6, and CLN8 are thought to occur in the ER, while CLN5 is believed to interact with TPP1 and CLN3 in lysosomes [Lyly et al., 2009]. Outside the NCL gene network, some NCL genes are implicated in various molecular pathways. For example, CLN3 is linked to endocytosis (via interactions with β-fodrin, Hook1, and Na-K+ ATPase complex), membrane trafficking (interacting with Hook1), autophagy, and apoptosis (through connections with Ca2+-dependent protein calsenilin, that mediates calcium-induced cell death) [Cao et al., 2006; Chang et al., 2007; Lane et al., 1996; Luiro et al., 2004; Uusi-Rauva et al., 2008]. No interaction partners have yet been identified for MFSD8 and CTSD. The disease mechanism can be further studied by crossing mice with mutations in NCL genes, such as the naturally occurring nclf and mnd mouse models, and those developed experimentally for Ppt1, Cln2, Cln3, Cln5, and Ctsd (http://www.ucl.ac.uk/ncl/mouse.shtml).

Diagnostics

Diagnosis of NCL relies upon a combination of clinical features and various laboratory examinations. EM of tissue for the distinct pathology, in combination with the typical clinical features, is confirmatory of NCL disease (Fig. 9). Enzyme or genetic data provide details of genetic type, with the age at presentation and rate of disease progression giving additional details. Enzyme analysis for PPT1, TPP1, and CTSD activities, diagnostic for CLN1, CLN2, and CLN10 diseases, should always precede mutation analysis, whatever the age of onset [Fig. 9; Fritchie et al., 2009; Sohar et al., 1999; van Diggelen et al., 2001]. Sometimes for the sake of time and economic efficiency, enzymatic analysis of PPT1, TPP1, and/or CTSD precedes even EM analysis, if CLN1, CLN2, or CLN10 disease, respectively is suspected. For each age group, NCL is most likely to be caused by mutations in certain genes, but all may need to be considered in older patients (Fig. 9). For example, in infants and young children CLN1 or CLN2 diseases are most likely, and enzyme tests should be performed first, before considering CLN10 disease or late infantile variants subtypes. If the disease starts in a school-age child with rapid visual failure as the presenting symptom, CLN3 disease should be suspected [Anderson et al., 2006; Santavuori, 1988]. In these cases, vacuolated lymphocytes should be looked for in the blood followed by screening for the common 1-kb deletion and/or screening of the entire coding region of CLN3. The late infantile variant subtypes (CLN5, CLN6, CLN7, and CLN8 diseases) are clinically indistinguishable and molecular genetic analyses of individual genes remain the only sure way to distinguish the different entities and categorize the patients accordingly. As most of the known mutations in these genes are “private,” the development of a routine diagnostic DNA-based test suitable for all patients is unlikely [Sharp et al., 2003]. However, screening of specific mutations can be prioritized in certain populations where local founder effects have been identified, such as the CLN6 p.Glu72X mutation in Costa Rican patients, the MFSD8 p.Thr294Lys in Roma patients, or the CLN8 p.Arg24Gly mutation in Finnish patients with EPMR [Gao et al., 2002; Kousi et al., 2009; Ranta et al., 1999; Wheeler et al., 2002].

Figure 9.

Summary of protocol for the diagnosis of NCL. Upon suspicion of NCL, identification of typical storage material is confirmatory of the diagnosis. In case PPT1, TPP1, or CTSD deficiencies are suspected, enzyme analysis should be performed. Upon confirmation of the deficiency whole-gene mutation analysis is recommended. In other cases, assuming a confirmed pathological diagnosis of NCL, mutation screening of the remaining NCL-causing genes should be performed in the order indicated, which is based on the frequency of mutations known in each NCL gene. However, there is no reason to rigorously prioritize an order for testing for CLN5, CLN6, CLN7, or CLN8 unless the ethnic background of the patient is consistent with one type (e.g., CLN5 disease in Finland). A subset of cases will still not be genetically diagnosed, and samples could be used for identification of novel NCL genes.

For families in which the underlying mutations have been identified or which are caused by defects in CLN1, CLN2 or CLN10, prenatal testing via an enzymatic test or a DNA-based test is possible [Berry-Kravis et al., 2000]. In other families such testing requires detection of storage inclusions on chorionic villus cells by EM.

Conclusions

Significant advances during the past two decades in the NCL field have seen the identification of eight human NCL genes, and experimental animal models for almost all subtypes of the disease have been developed. The strategy required to reach a genetic diagnosis for each family affected by NCL disease has been clarified. Together with the 91 new mutations reported here, over 360 changes are now known to be associated with NCL disease. Mutation detection has not only benefited families by providing a molecular genetic diagnosis, but has also allowed genotype–phenotype correlations to be considered. Nevertheless, despite this progress made in recent years, the challenge of completing the genetic spectrum underlying this group of disorders remains. Toward this, new and more streamlined and high throughput approaches that are being developed are expected to speed genetic diagnosis even further in the next few years. Identification of the remaining NCL-associated genes will help to shed light on the disease mechanisms and will provide further insights for the development and design of novel therapies. Evaluation of future therapies however, is likely to be challenged by the complex genotype–phenotype correlations emerging in NCLs, as it is likely that modifying genes play a key role in disease manifestation. In future years it will be possible to test whether changes in a second NCL or other genes modify a genetically identified NCL phenotype, perhaps explaining the great inter- and intrafamilial variation that characterizes this group of disorders.

Acknowledgements

We thank all the research scientists, physicians, and diagnostic laboratories that regularly publish or submit their mutation data to the NCL mutation database (www.ucl.ac.uk/ncl/mutation)-–their contributions are acknowledged in the database. We also thank those who sent us summaries of their existing or new mutation data or pointed us in the right direction to obtain this data, including: Monique Losekoot of Laboratory for Diagnostic Genome Analysis, Leiden University Medical Centre, The Netherlands; Sam Loughlin, Regional Molecular Genetics, Great Ormond Street Hospital, London, UK; Hannah Mitchison, Institute of Child Health, UCL, London, UK; Gabriela Storkanova, Inst. Inherited Metabolic Disorders, Charles University and General University Hospital in Prague, Czech Republic; Filippo Santorelli, Molecular Medicine and Neurology, IRCCS Bambino Gesà Hospital, Rome, Italy; Montserrat Milà Recasens, Cap de Secció de Genètica Molecular Servei de Bioquímica i Genètica Molecular Hospital Clínic, Barcelona; Inés Noher de Halac, Centro de estudio de las metabolopatias congénitas (CEMECO)-Hospital de Niños, Córdoba, Argentina; David Palmer, Lincoln University New Zealand; Thomas Muller and Adam Coovadia, of the Molecular Genetics Laboratory, All Children's Hospital, St Petersburg, USA; Francisco Laranjeira, Biochemical Genetics Unit, Medical Genetics Center, National Institute of Health, USA; Winnie Xin, Neurogenetics DNA Diagnostic Laboratory, Mass General Hospital, USA. We thank Isa Uski for excellent technical assistance. We would like to thank Hannah Mitchison and Marjo Kestilä for critical comments that helped improve the quality of this mutation update. MK is a Biomedicum Helsinki Graduate School fellow.

Ancillary